The case for limiting government semiconductor subsidies

Recently, several U.S. auto plants have had to shut down due to a shortage of microchips. Based on a concern about disruptions to supply chains for critical inputs like these, the Senate passed the U.S. Innovation and Competition Act (USICA) with bipartisan support. The bill, which is now being considered by the House of Representatives, includes several provisions to promote key U.S. industries, including semiconductors. 

One part of the legislation, the CHIPS Act, allocates $52 billion to subsidize the construction of new chip factories in the United States. Officials believe it could result in seven to 10 new U.S. factories. Although a national security case can be made for it, too much subsidy can do more harm than good.

Part of the motivation is to reduce the reliance of U.S. firms on imported microchips. A growing share of semiconductors are imported, with only 12 percent produced domestically, down from 37 percent since 1990. 75 percent are manufactured in Asian manufacturing plants. 

Proponents defend the subsidies as necessary to protect the supply chain for semiconductor chips, which are used by the military and many other U.S. industries. Taiwan is a major source of computer chips, particularly those used by the auto industry. But many U.S. officials believe the chance of a conflict between China and Taiwan has grown. Such a conflict could disrupt the supply of chips. 

The current chip shortage was not caused by any disruption in shipments from other countries. Many automakers canceled their orders after car sales declined sharply in March of 2020. In response, fabricating plants cut production and increased their production of chips for consumer electronics, such as laptops and games, whose sales have been surging. They earn higher margins on the latter, which has made it that much more difficult to ramp up production of the chips needed by automakers when car sales unexpectedly rebounded before the end of 2020.

Increasing domestic production capacity is not the only way to reduce the risk of supply chain disruptions. In some cases, it may be less costly to stockpile less-expensive imported goods rather than produce them domestically. Automakers that did not stockpile chips have been largely shut out of the market in recent months. Toyota, which set up a stockpile, did better than most.

Additionally, when supply is temporarily disrupted, rising prices give consumers an incentive to delay the purchases of nonessential items that use semiconductors. That helps the chips we have go toward the most important uses.

Another common argument for subsidizing domestic production is to make sure American workers have the expertise to produce domestically when market conditions warrant. But the United States already produces a substantial quantity of semiconductors — just not enough to meet domestic demand. 

There is a considerable risk that in response to current shortages, too many chip foundries will be built, both in the United States and elsewhere. This is doubly true if governments overreact. The result would be a glut of semiconductors. If prices drop because supply grows more rapidly than demand, it may mean some producers will need future subsidies just to survive.

Without the subsidies, firms would be more cautious about building or expanding foundries. If long-term production capacity is truly insufficient, high prices and anticipated profits give firms the right incentives to build or expand and satisfy demand at cost-covering prices.

We should not be cavalier in how we treat this industry. The integrated global supply chain of semiconductors has resulted in enormous innovation, steadily declining costs and performance improvements that play a critical role in the growth and development of information technology. U.S. firms have specialized in R&D intensive activities, such as designing semiconductors, and firms in other countries, such as Korea and Taiwan, have specialized in manufacturing them. Subsidies to encourage more foundries to locate in the United States should be limited in order to keep each nation doing what it does best.

Politicians will inevitably compete to get subsidies for foundries in their states or districts, regardless of what’s most cost-effective. If subsidies result in plants built where production costs are higher than otherwise, the result could be slower U.S. economic growth. If we lean too heavily into policies that lower growth, we can even undermine our ability to maintain and improve military preparedness.

Providing subsidies to expand U.S. chip production capacity may make sense when a leading supplier such as Taiwan is vulnerable to attack by China. But it is important to continue to import semiconductors from a diverse group of low-cost foreign suppliers, particularly those of trusted allies.

Tracy C. Miller is a senior policy research editor with the Mercatus Center at George Mason University.